Determination of Phenol and m-Cresol in Complex Biochemical Mixtures

by distillation at pH 8.0 to 8.5, and the sum of phenol and m-cresol is determined by the sulfanilic acid method. Phenol is then separated from the cr...
1 downloads 0 Views 545KB Size
Determination of Phenol and m-Cresol in Complex Biochemical Mixtures Ap p 1ica t io n of Co un t e rc ur r e n t D is t r i b ut io n Method BENJ. WARSHOWSKY AYD E. J. SCH4NTZ WITH TECHNICAL ASSISTANCE OF \IILLIRD V.RICE Camp Detrick, Frederick, M d . Phenol can be quantitatively determined in the presence of the isomeric cresols; mixtures of m-cresol and phenol can be determined in complex biochemical samples. The phenolic compounds are removed from the interfering substances by distillation at pH 8.0 to 8.5, and the sum of phenol and rn-cresol is determined by the sulfanilic acid method. Phenol is then separated from the cresols by submitting an aliquot of the distillate to a 24-plate countercurrent distribution. From the amount of phenol found in a selected series of tubes, the concentration in the original sample can be calculated. rn-Cresol, in the presence of phenol, is determined by difference by subtracting the value for phenol from the total phenolic content of the sample. Some volatile phenolic compounds may interfere with the method.

A

LTHOUGH numerous procedures have been described for the determination of phenol or phenolic compounds in simple aqueous solutions ( I S ) , no satisfactory method for the simultaneous determination of dilute solutions of mixtures of these compounds in biological samples is available. Such a procedure would undoubtedly have appreciable value for clinical, bacteriological, and industrial analysis where it is often desirable to determine the concentration of each component of a phenolic mixture. Several procedures have been applied to this problem, but they have only a limited scope. Dull (5) and Potter and Williams (9) determine o-cresol in phenol by forming a complex with cineole and determine the complex by freezing point data. Another method ( I d ) is based upon the lowering of the freezing point of phenol by cresol. .I similar procedure, but one that has been found to yield more reliable results, is the cloud point method of Seaman, Sorton, and Foley ( l a ) . These methods are, however, applicable only to simple mixtures of phenol and cresol where the only other coinponent is water. h different approach to the problem was described by Bielenberg and Fischer ( I ) , who coupled mixtures of hydroxybenzene compounds, such as phenol and cresol, with diazotized p-nitroaniline to form individual azo colors which were then separated chromatographically and determined colorimetrically. Miller and Urbain (8) report that phenol is quantitatively oxidized by chromic acid while cresols and the higher phenols such as resorcinol are unaffected. They propose a method for estimating phenol by difference before and after chromic acid oxidation. This procedure was investigated by the present authors and i t was found that some cresol iyas oxidized as well as the phenol. A spectrographic method has also been used for determining the concentration of phenol and cresol in mixtures ( I O ) , but t'his procedure is not applicable t o dilute aqueous solutions of these compounds. Ina3much as the original purpose of thiq investigation was t'he determination of the phenol cont'ent of baet'eriological materials and media, several phenol methods were tried directly on the complex samples. S o one method, however, was found satisfacfor this purpose. The solution of this phase of the problem eventually found in the removal of the phenol from the bacteriological materials that contained the interfering substances. This separation was accomplished by distilling the phenol from the mixture after adjusting the pH to 8.0 to 8.5. -4t this p H the interfering substances were held back as salts while the phenol was free to distill. I t was found that when 80 to 90Vo of the water was

distilled, 99 to 1007, of the phenol passed over into the diitillatr. After the separation of the phenol from the interfering substances, the determination could be completed by any one of the standard procedures. Although the Folin and Ciocalteu method (6') was initially used for the colorimetric determination of the phenol, it was not so sensitive as the diazotized sulfanilic arid method. Consequently, the latter procedure was followed, as it n as desired to determine accurately microgram quantities of the phenolic compounds. Once it was learned that phenol could be quantitatively separated from complex biochemical substances, the effect of cresol was investigated. This study shoaed t h a t cresol and phenol behaved similarly and could not be easily distinguished from each other. It \vas eventually decided to investigate the possibility of separating phenol from cresol by means of the countercurrent distribution technique of Craig and eo-workerb. For a compreheniive description of the method, apparatus required, theoiy involved, and some applications of the technique, it is recommended that the papers of Craig ( 2 , 3, 4, 7, 11, 15) be consulted. The fundamental basis of countercurrent distribution is the partition of a soluble substance between two immiscible phases. The relative concentration of the solute in each of the phases is a function of the partition coefficient of the substance, and, under specified experimental conditions, is a constant ratio. This partition coefficient is an identifying characteristic of pure compounds and has been used t o determine the heterogeneitv of otherwise inseparable miutures. This was accomplished by utilizing the fact that although compounds may be chemically similar in structure, they often may be shown t o have sliglitlv different partition coefficients when distributed between a proper choice of two immiscible yolvents. This slight difference in the preferential solubility in one of the qolvents can be amplified by repeating the procrss a sufficient number of times, so that for practical purposes a solution may be obtained in m-hich onl! one component is present; the more soluble compound may be estracted, or all but the least soluble component of the mixture removed. The Craig-machine ( 3 ) is a convenient apparatus for conducting this series of stepwise liquid-liquid distribution meawrementa without necessitating the tedious and cumbersome use of a large number of individual separatory funnel extractions. 41though it was successfully utilized in the resolution and characterization of the penicillins, the isolation of antibiotic principles from Aspergzllus ustus, and the study of the homogeneity of 951

952

ANALYTICAL CHEMISTRY

antimalarials, it has received only limited use in the field of analytical chemistry. This instrument promises to be of appreciable value in the solution of analytical problems in which it is difficult to eliminate the error due to interfering substances. That it is also applicable to simultaneous determination of a mixture of chemically similar compounds w m shown by Sato, Barry, and Craig ( I I ) , who succeeded in separating and quantitatively estimating within 2 to 3% the volatile normal fatty acids (C, to C,) in a mixture. The procedure was equally successful in the present problem of resolving the components of the phenolcresol mixture sufficiently to permit quantitative determination of each constituent. This was ultimately effected by subjecting the mixture to a 24-plate distribution between carbon tetrachloride and potassium acid phthalate buffer of pH 5.0. From the amount of phenol detected in one of the phases of a selected series of tubes, the quantity of phenol present in the original sample was calculated. m-Cresol was found by difference by subtracting the value for phenol from the total amount of phenol plus ?+cresol in the initial mixture.

In this manner, all the cells of the lower section can be filled in one operation by pouring solvent into a single tube. As the tubes are interconnected, the liquid will overflow into all the holes. To ensure the presence of equal volumes of solvent in all tubes, the upper section is removed and the remaining sections are filled so that the meniscus is level with the surface of the bottom section. The top section is then replaced and set so that the numbered Cubes of both sections correspond, and exactly 8 ml. of the buffer (previously saturated with carbon tetrachloride) are added by means of a buret to all the tubes except the zero tube. An aliquot of the sample containing from 100 to 150 micrograms of both phenol and m-cresol is placed in the upper part of the zero tube and sufficient buffer to bring the total volume of the upper layer to 8 ml. is added.

9-

04-

u

s

$

03-

L

APPARATUS AND REAGENTS

Stock standard solutions of phenol and m-cresol were prepared by diluting Merck’s pure analytical reagent grade phenol and synthetic m-cresol (prepared from m-toluidine), boiling point 201 ’ C., with distilled water to give a concentration of 1.000 mg. per ml. These solutions were diluted further to 5 micrograms per ml. for use in preparing the standard curve for the colorimetric determination of phenol and cresol. The sulfanilic acid reagent was prepared according to the procedure described by Snell (IS): Mix 4.5 grams of sulfanilic acid with 45 ml. of concentrated hydrochloric acid and dilute to 500 ml. with water. In another 500-ml. volumetric flask dissolve 25 grams of sodium nitrite in water and dilute to the mark. Cool 3.00 ml. of the sulfanilic acid solution in a 100-ml. volumetric flask in an ice-water bath, add 3.00 ml. of the sodium nitrite solution, cool for 5 minutes, and then add 12.0 ml. of the sodium nitrite solution. Make up to a volume of 100 ml. with distilled water and keep the final reagent solution cold. This reagent can be used after 15 minutes, and is prepared fresh daily. The two immiscible solvents used for the countercurrent distribution of the phenol arid m-cresol were Merck’s reagent grade carbon tetrachloride and 0.1 ,?I potassium acid phthalate buffer of pH 5.0. For the separation of the phenolic compounds from the interfering substances in the sample, a simple Iijeldahl distillation unit, con4sting of a 100-ml. Kjeldahl flask, a water-trap, and a condenser was used. The Craig apparatus was a 24-tube, stainless steel unit, which was purchased from Otto Post, Maspeth, Y. Y., and had a capacity of 8 ml. for each tube of the lower section. A Coleman Universal spectrophotometer, Model 91, was utilized for colorimetric determination of the phenol. Matched cuvettes of 18-mm. diameter were wed for measuring the intensity of light transmitted. ANALYTICAL PROCEDURE

Distillation of Phenolic Mixture. In the present work it was found convenient to employ a sample containing approyimately 1 mg. of phenol and 1 mg. of m-rresol. Samples with a higher concentration of the phenolic compounds can also be submitted to the distillation procedure if smaller aliquoLs of the distillate are subsequently used for the countercurrent separation. Approximately 25 ml. of 0.1 M sodium bicarbonate solution a t p H 8.1 were added to the sample in a 100-ml. Kjeldahl flask and

the total volume was diluted to about 60 ml. with distilled water. The flask was gently heated and the distillate was collected in a 200-ml. volumetric flask immersed in an ice bath. The distillation was continued until about 2 to 5 ml. of solution remained in the Kjeldahl flask. The end of the condenser was rinsed with water and the distillate diluted to the 200-ml. volume. Separation of Phenol from rn-Cresol. The carbon tetrachloride which was used as the heavy phase in the distribution apparatus was saturated with the aqueous buffer by shaking 250 ml. with 250 ml. of the phthalate buffer. The lower half of the apparatus is filled with the carbon tetrachloride in the manner prescribed by Craig, which consists of adjusting the upper section so that each hole covers half of the corresponding hole in the lower section.

:p: 0 2 OI-

400

450

550

500 WAVE

LENGTH, m

500

p

Figure 1. Absorption Spectrum of Colored Product Resulting from Reaction of Diazotized Sulfanilic Acid with Phenol and m-Cresol

The apparatus is assembled, with all the parts in place and water-tight. With the holes arranged in the starting position, the instrument is rotated 25 times a t a speed which will ensure proper mixing of the phases. This can be observed in the glass indicator tube that is attached to the metal body of the machine and contains 8 ml. of both liquids. When equilibrium is attained the layers are allowed to separate, and the upper section is carefuily moved so that the upper portion of the zero tube is now superimposed on the lower half of the S o . 1 tube. The same series of manipulations is repeated for all 24 tubes. Upon completion of the partition operations, the pressure plate is removed and a 4-ml. aliquot of the water layer in each of the tubes is pipetted out into a test tube in preparation for the colorimetric determination of the phenolic content of the individual tubes. In the actual determination of an unknown sample, the analysis of only three or four suitably selected tubes is necessary. Colorimetric Dete:mination. T o the aliquot from the countercurrent distribution treatment are added 1 nil. of a 2OYc sodium carbonate solution and water to make the total volume 8 ml. Two milliliters of the cold sulfanilic acid reagent are added to the mixture, which is shaken for 1 to 2 minutes, and the per cent tranqmittance is read in the spectrophotoxeter against blank of the reagents at a wave length of 425 inp. Tne amount of total phenols in the aliquot sample is calculated (as phenol) from a standard curve constructed from a series of known amnunfs of phenol in the range of 1 to 20 micrograms. CALCULATIONS

The calculations for the amount of phenol ill the aliquot ~aiiiplr subjected to the countercurrent distribution are based on the theoretical equations for the partition of a substance between two immiscible solvents. These equations have been discussed by Williamson and Craig (15), and involve a consideration of the partition coefficient of phenol in the system selected and the number of single extractions used to effect the desired separation. It can be shown experimentally that the solute in such a system is distributed between the two phases in accordance with the blnomial expansion theorem. Consequently, by determining the amount of solute in any particular tube, it is relatively simple to calculate the concentration in the original sample. For reasons mentioned below, it was found more accurate to estimate the initial phenol concentration from the phenol content of the aqueous layer only rather than from the total amount of the entire

V O L U M E 20, NO. 10, O C T O B E R 1 9 4 8

953

tube. Thus, merely by determining the weight of phenol in the water layer of three or four tubes, it is possible to calculate the concentration of phenol in the original sample.

Table 11.

Effect of pH on Distillation of Phenol and rn-Cresol Per Cent Recovery

The expression relating the partition coefficient with the fraction of the original substance present in any tube, r, in a distribution of n transfers or plates is given by Craig as:

PH

Phenol

8.0 8.5

100

Tr7

~ l i e l - eP is the factor for relttting T, to T,+ and T , and I are amount of solute in the t u b e r and r 1. The factor, F , is, in turn, calculated by the formula,

-

F= n f l - r r

100

100

101

94.0 80.6 61 .O

9.0 Y .5

10.0 where T,,,= fraction of original substance in tube r, n = total number of transfers, and k = partition coefficient. The method used in this study for the determination of the partition coefficient of the solute utilized the condition that the ratio of the concentrations in any two adjacent tubes is a specific function of the partition coefficient. This relationship is given by the equation

m-Cresol 96.2 89.7 74.4

from the interfering substances. The results of t!iis work showed that recoveries of almost 100% phenol can be achieved (Table I). If the pH of the bacteriological sample x a s not previously adjusted to 8.0 to 8.5, errors resulting from interfering substances aniounted t o as much as 2OT0 relative. .A similar series of esperiinents n-ith cresol gave essentially the same results. In an attempt to separate cresol and phenol by distillation at different pII values mixtures of the two were distilled from aqueous solutions at, various pH's ranging from 8 to 10. Table I1 shons the effect of pH on the recovery of these compounds by distillation. The values obtained in these esperiments make it clear that distillation at a controlled pII is unsatisfactory as a rn+an* or scqmrating phenol from the crew!s,

Thus, by selecting any two adjacent appropriate tubes (those tubes corresponding to the peak of the distribution curve) the partition coefficient can be readily calculated from the amount of phenol present in the aqueous layer. For example, in one experiment involving a 24-plate transfer of phenol and m-cresol, tube 9 contained 15.2 micrograms of phenol in the upper water layer and tube 8 contained 16.2 microerams of Dhenol. For tube 9. there24+ 1 - 9 16 - -. fore, F = I

9

9

The partition coefficient, k =

Y lfi --

1525

X n*

- - .- -

= 0.529

By substituting in Equation 1 the correct value for k and expanding the expression to find the rth term, the fraction of the original phenol present in any tube can be obtained. For example, using the sanie data as in the preceding paragraph for tube

OZl 0.1

i -

10

24

9, T,,,r = (1.30 X lo8) X ( m 51 s 3 ) X (0.529)g = 11.30 X 106) X (3.85 X 10-6) X (3.25 X = 0.164. By dividing this fraction into the value for the uhenol content found in the aqueous layer of tube 9, the phendl present in the sum of all the 15 25 water layers will be obtained-e.g., - = 93.0 micrograns. 0.164 Inasmuch as this figure represents only the quantity found in one phase-i.e., the water layers-it is necessary to calculate from the partition coefficient the amount present in the other solveni layer. Because, in the present instance, the partition coefficient is 0.529, the amount of phenol present in the carbon tetrachloride is 0.529 X 93.0 = 49.3 micrograms. The sum of these t15-o values (93.0 49.3 = 142.3) gives a quantitative measure of the phenolic content of the aliquot sample submitted to the countercurrent distribution aeparation.

+

Table I. Phenol Added Mg. 1000 800 500 200 100 2.00 1 .oo 0.00

Recovery of Added Phenol Phenol Fouad

Ma. 1000 790 500 200

100 1.Y8 1.01 0.00

Phenol Recovered

% 100 98.8 100 100 100 99.0 101

DISTILLATION AND DETERMINATION OF PHENOLIC MIXTURES

Using the conditions described under Procedure, samples of complex biochemical mixtures containing various amounts of phenol were used to investigate the reliability of the separation

// 15

20

PHENOL, H I C R O G R I Y S

Figtire 2.

Efl'ect of Concentration of Phenol

(it1

Intensity of Color .1partial in>eestiyatioiiof the ~ p a ~ i h c * iui ' \ t h d;btillatioll PI"cedure was also undertaken. Ar the only substances that i t i)ulir intri iere with the method as described are ~ o l a t i l ephenolic coil]pounds, higher phenols were subjected to the same treatment Ir was found that orcinol, resorcinol, phloroglucinol, and hydroquinone were not distilled under the conditions used and did not affect the accuracy of the method; 2-hydroxy-1,4-dimethy! phenol, however, did interfere. No attempt was made to adapt the present procedure to mixtures containing this compound. The diazotized sulfanilic acid method for determining the phenol and m-cresol was found very satisfactory. As no curve of the absorption spectrum of the colored product was available, this information was determined. On the basis of the results illustrated in Figure 1, it was considered advisable to select 425 mp as the wave length for all subsequent colorimetric measurements. The relationship between the concentration of phenol and the log of per cent transmittance a t this wave length is shown in Figure 2 Except for differences in molecular weight, m-cresol was found to give the same wansmittance as phenol. Therefore, in the determination of the total phenolic content of a sample, only one standard curve for phenol is required; m-cresol, obtained ultimately by difference, will be computed in terms of phenol and must be corrected by multiplying by the ratio of the molecular weights. The intensity of the color developed in an aqueous medium was found to be entirely different from that found in carbon tetrachloride. For this reason, all determinations were

ANALYTICAL CHEMISTRY

954

layer of several tubes showed that tube 3 contained the maximum amount. The average partition coefficient was calculated for Phenol Phenol Phenol tubes 2, 3, and 4 and found to be 0.447. Using this average value Present Fqund Calcd. as the most probable, calculations similar to those shown in the In in Partition in Run Original Tube Buffer CoefOriginal preceding section for computing the concentration of phenol in Xo, Sample So. Layer ficient Sample Recovery the original sample were made. The data of this and another 218. Mg. Mg. % 1 140 0 2 2 similar experiment are shown in Table 111. The initial indication of the probability of separating phenol from the cresols was the observation that the partition coefficient of rn-cresol was significantly different from that of phenol under the conditions used Partition coefficient measurements with m-cresol between the U b 0 solvents selected gave an average value of 1.25. As a mavimum 4 0 degree of separation was desirable, the countercurrent distribution Av. 0447 140.1 100 1 2 140 0 2 R mas extended to a 24-plate transfer, and applied directly to a syn16.9 thetic mivture of the two compounds. Figure 3 is a graphic repre22.4 0.456 148.0 27.0 0.446 148.0 sentation of the distribution curve obtained upon analysis for total 21.3 0.451 148 2 11.6 phenols of the buffer layer of each tube. .ilso included in this 3 4 figure is a plot for the theoretical curve of phenol calculated 0 0 from the equation for the binomial expansion theorem, as indi0 cated previously. The results of the actual determination of the .\v. 0.451 148 0 105.8 .lv., runs 1 and 2 144.8 102.9 phenol and m-cresol content of several samples of a synthetic Table IV. Separation and Determination of Phenol a n d m-Cresol misture are given in Table IV. Total Total Phenol Content Calculated m-Cresol The samples were from a sludge Saniple Phenol m-Cresol Phenols from Tube ( b y Difof B. globigii to which known No Present Present Found" S o . 6 No. 7 S o . 8 S o . 9 Av. Error ference) Error Y Y Y Y Y Y / Y n Y % amounts of phenol and rn-cresol 134 0 -.5.0 rrere added. They were sub1 140.0 141.0 262.0 139.Y 136.6 150.0 1 4 6 . 6 143.0 f2.1 2 140.0 141.0 263.6 139.2 1 3 7 . 9 138.3 142.9 1 3 9 . 6 139.8 0 . 9 -0.3 89.5 -5.8 jectedtotheentireprocedureJin3 105.0 95.0 185.8 104 3 104.5 105,8 111.2 1 0 6 . 4 f1.4 eluding distillation and counter' Phenol and m-cresol determined in terms of phenol. current distribution separation. Considering complex nature of the material analyzed, errors indicated in Table IV are not excessive. LEGENO - - _ EXPERIMENTAL Although the procedure has been described for the determinaTHEORETICAL tion of phenol and the meta form of cresol, the phenol can be w separated from and determined directly in the presence of the \ other isomers as well. However, since the cresol is eventually determined by difference, it would be necessary to investigate the effect of all the pure synthetic isomers on the intensit,y of the colorimetric measurements before mixtures of all the cresols can be accurately determined. This study has not yet been undertaken. Table 111. Determination of Phenol in Known Samples

/

ACKNOWLEDGMENT

The authors wish to acknowledge the assistance of J. P. Crispell in investigating the applicability of many phenol methods and of Ilse Schwab in performing some of the chemical determinations. TUBE

NUMBER

Figure 3. 24-Plate Countercurrent Transfer Distribution of Phenol and rn-Cresol between Carbon Tetrachloride and Phthalate Buffer 140 Y each. pH 5.0. Broken line represents experimental curve for 140 y each of phenol and m-cresol. Solid line is theoretical curve for phenol only

LITERATURE CITED (1) Bielenberg, T.,and Fischer, L., Brennstof-Chem., 22, 278 (1946). (2) Craig, L. C., J . B b l . Chem., 150, 3 3 (1943). (3) Ibid.. 155, 519 (1944). (4) Craig, L. C., Golumbic, C., Mighton, H., and Titus, E., J . B i d . Chem., 161, 321 (1945). (6) Doll, H., Arch. Pharm., 274, 283 (1936). (6) Folin. 0..and Ciooalteu. V.. J . Biol. Cham.. 73. 627 (1927).

(7j Hopeboom, G. made on only the aqueous layer of the tubes and not the total contents. COUh-TERCURRENT SEPARATIOh

I t was found that the partition coefficient of phenol is not markedly affected by the pH of the aqueous phase. Severtheless, as a precaution, a buffer solution of sufficient buffer capacity to maintain a constant pH was used. A buffer of about pH 5 was arbitrarily selected on the basis that the partition coefficient of phenol a t this pH was sufficiently different from the partition coefficient for rn-cresol to make the separation possible, and yet was not too far removed from the optimum value of 1. A preliminary application of the procedure to the quantitative determination of a known amount of phenol was made using a 10plate transfer. Determination of the phenol content of the buffer

H., and' Craig, L. C., J . Roc&feller Inst. M e d . Res., 129, 8 9 (1946). (8) Miller, J. N., and Urbain, 0. M., IND.ENG.CHEM.,ANAL.ED., 2, 123 (1930). (9) Potter, F. M.,and Williams, H. B., J . Soc. Chem. Ind.,51,591 (1932).

(10) Robertson, 77'. W., Ginsburg, N., and Matsen, F. A.,IND. ENO. CHEM., -4NAL. ED., 18, 746 (1946j. (11) Sato, Y., Barry, G. T., and Craig, L. C., J . Biol. Chem., 170, 501 (1947).

Seaman, W., Norton, A . R., and Foley, R. T., IND. ENG.CHEM., AKAL.ED., 15, 159 (1943). (13) Sne!l, F. D., and Snell, C. T., "Colorimetric Methods of Analpsis," Vol. 11, pp. 348-76, New York, D. Van Nostrand Co.,

(12)

1937. (14)

Weiss, J. M., and Downs, C. R., J . Ind. Eng. Chem., 9, 569

(15)

Williamson, B., and Craig, L. C.. J . Biol. Chem., 168, 687 (1947).

(1917).

RECEIVED February 21, 1948.

,